HERWIG is a general-purpose Monte Carlo event generator, which includes the simulation of hard lepton-lepton, lepton-hadron and hadron-hadron scattering and soft hadron-hadron collisions in one package. It uses the parton-shower approach for initial-and final-state QCD radiation, including colour coherence effects and azimuthal correlations both within and between jets. This article updates the description of HERWIG published in 1992, emphasising the new features incorporated since then. These include, in particular, the matching of first-order matrix elements with parton showers, a more correct treatment of spin correlations and heavy quark decays, and a wide range of new processes, including many predicted by the Minimal Supersymmetric Standard Model, with the option of R-parity violation. At the same time we offer a brief review of the physics underlying HERWIG, together with details of the input and control parameters and the output data, to provide a self-contained guide for prospective users of the program. This version of the manual (version 3) is updated to HERWIG version 6.5, which is expected to be the last major release of Fortran HERWIG. Future developments will be implemented in a new C++ event generator, HERWIG++.The program is written in Fortran and the user has to modify the main program HWIGPR to generate the type and number of events required. See section 8.1 for a sample main program. The program operates by setting up parameters in common blocks and then calling a sequence of subroutines to generate an event. Parameters not set explicitly in the block data HWUDAT or in HWIGPR are set to default values in the main initialisation routine HWIGIN. Output data are delivered in the LEP standard common block HEPEVT [25,26]. Note that all real variables accessible to the user, including those in HEPEVT, are of type DOUBLE PRECISION.Since version 6.3, to take account of the increased energy and complexity of interactions at LHC and future colliders, the default value of the parameter NMXHEP, which sets the array sizes in the standard /HEPEVT/ common block, has been increased to 4000.To generate events the user must first set up the beam particle names PART1, PART2 (type CHARACTER*8) and momenta PBEAM1, PBEAM2 (in GeV/c), a process code IPROC and the number of events required MAXEV.See section 4 for beams and processes available. All analysis of generated events (histogramming, etc.) should be performed by the user-provided routines HWABEG (to initialise the run), HWANAL (to analyse an event) and HWAEND (to terminate the run).A detailed event summary is printed out for the first MAXPR events (default MAXPR = 1). Setting IPRINT = 2 lists the particle identity codes, properties and decay schemes used in the program.The programming language is standard Fortran 77 as far as possible. However, the following may require modification for running on some computers • Most common blocks are inserted by INCLUDE 'HERWIG65.INC' statementsthe file HERWIG65.INC is part of the standard program package.• Many common blo...

In this paper we describe Herwig++ version 2.2, a general-purpose Monte Carlo event generator for the simulation of hard lepton-lepton and hadron-hadron collisions. A number of important hard scattering processes are available, together with an interface via the Les Houches Accord to specialized matrix element generators for additional processes. The simulation of Beyond the Standard Model (BSM) physics includes a range of models and allows new models to be added by encoding the Feynman rules of the model. The parton-shower approach is used to simulate initial-and final-state QCD radiation, including colour coherence effects, with special emphasis on the correct description of radiation from heavy particles. The underlying event is simulated using an eikonal multiple parton-parton scattering model. The formation of hadrons from the quarks and gluons produced in the parton shower is described using the cluster hadronization model. Hadron decays are simulated using matrix elements, where possible including spin correlations and off-shell effects.

We present a new general algorithm for calculating arbitrary jet cross sections in arbitrary scattering processes to next-to-leading accuracy in perturbative QCD. The algorithm is based on the subtraction method. The key ingredients are new factorization formulae, called dipole formulae, which implement in a Lorentz covariant way both the usual soft and collinear approximations, smoothly interpolating the two. The corresponding dipole phase space obeys exact factorization, so that the dipole contributions to the cross section can be exactly integrated analytically over the whole of phase space. We obtain explicit analytic results for any jet observable in any scattering or fragmentation process in lepton, lepton-hadron or hadron-hadron collisions. All the analytical formulae necessary to construct a numerical program for next-to-leading order QCD calculations are provided. The algorithm is straightforwardly implementable in general purpose Monte Carlo programs.

We propose a version of the QCD-motivated`k ? ' jet-clustering algorithm for hadron-hadron collisions which i s i n v ariant under boosts along the beam directions. This leads to improved factorization properties and closer correspondence to experimental practice at hadron colliders. We examine alternative denitions of the resolution variables and cluster recombination scheme, and show that the algorithm can be implemented eciently on a computer to provide a full clustering history of each e v ent. Using simulated data at p s = 1 : 8 T eV, we study the eects of calorimeter segmentation, hadronization and the soft underlying event, and compare the results with those obtained using a conventional conetype algorithm.